AN OPTICAL OBJECT DETECTION SYSTEMFOR SENSING OBSTRUCTIONSTO LOW SPEED VEHICLES
George Alfred Burman
NAVAL POSTGRADUATE SCHOOL
Monterey, California
THE!51SAN OPTICAL OBJECT DETECTION SYSTEM
FOR SENSING OBSTRUCTIONSTO LOW SPEED VEHICLES
by
George Alfred Burman
Thesis Advisor G. L. Sackman
June 1972
T14851?
Approved fan. pubtic. naZtase,; dLi&Ubuutlon antonitzd.
An Optical Object Detection Systemfor
Sensing Obstructions to Low Speed Vehicles
by
George Alfred BurmanLieutenant, United States Navy
B.S., California State Polytechnic College, 1963
Submitted in partial fulfillment of the
requirements for the degree of
MASTER OF SCIENCE IN ELECTRICAL ENGINEERING
from the
NAVAL POSTGRADUATE SCHOOLJune 1972
ABSTRACT
An object detection system to sense obstructions in the path of
low speed vehicles is described. The system uses a pulsed GaAs
diode laser as an illuminator, and a PIN photodiode as a detector.
Reliable detection of objects with an effective area as small as
20.5 ft was achieved. A signal processor using active filters and a
phase-locked loop tone decoder was employed for both phase and
frequency rejection of undesired signals, and detection of objects under
high ambient light conditions.
TABLE OF CONTENTS
I. INTRODUCTION '
7
II. DETAILED SYSTEM DESCRIPTION 13
A. TRANSMITTER 13
1. Laser Pulsing Considerations 13
2. SCR Pulsing Circuit 14
3. Trigger Circuit 18
4. Charging Circuit 18
5. Optics-- 19
6. Safety Considerations 20
B. RECEIVER 20
1. Optics 20
2. Photodiode Considerations 22
3. Impedance Converter 23
4. Preamplifier 25
5. Power Supply 26
C. SIGNAL PROCESSING SECTION 26
1. General 26
2. Active Filter 2 7
3. Phase-Shifter, Buffer, and Limiter . 31
4. Tone Decoder 32
III. TEST RESULTS 35
A. RANGE TESTS 35
1. Reflection Tests 35
2. Target Response 37
3. Direct Line_of-Sight Tests 37
B. SPURIOUS RESPONSES 41
1. Ambient Light 41
2. Dust and Smoke 41
3. Electromagnetic Compatability 42
4. Adjacent Transmitter Interference 42
IV . CONCLUSION 44
A. RANGE RESPONSE 44
B. TARGETS 45
C . ENVIRONMENTAL CONDITIONS 45
APPENDIX A Photographs of Various Voltage andCurrent Waveforms 46
APPENDIX B Active Filter Frequency Response 49
BIBLIOGRAPHY 50
INITIAL DISTRIBUTION LIST 51
FORM DD 1473 52
LIST OF DRAWINGS
1. Desired Detection Zone 8
2. Detection Zones Formed by Transmitter-Receiver Pairs 10
3. Transmitter-Receiver Block Diagram 11
4. Normalized Threshold Current vrs. Junction Temperature 15
5. Transmitter Schematic 16
6. Transmitter Beam Optics 19
7. Receiver Optics 21
8. Photodiode Impedance Converter and Preamplifier Schematic - 24
9. Photodiode and Impedance Converter Equivalent Circuit 23
10. Signal Processor Block Diagram 28
11. Spectral Content of Fluorescent Lights 30
12. Active Filter Schematic 29
13. Phase -Shifter, Buffer, and Limiter Schematic 31
14. Tone Decoder Schematic 32
215. Range Response of a 1 ft Target 36
16. Direct Line-of-Sight Range Response 40
17. Radiated Interference Spectrum 43
ACKNOWLEDGEMENTS
The author wishes to thank Professor George L. Sackman, for his
assistance as Thesis Advisor, and for the original formulation of the
system concept; the Hewlett Packard Corporation and the Signetics
Corporation for providing components and technical assistance; and
his wife, Martha, for typing and assistance in taking data.
I. INTRODUCTION
An object detection system was required for low speed cargo
vehicles, similar to fork-lifts, to prevent personnel and equipment
injury. The vehicles operate in a long warehouse corridor, and are
programmed for automated loading and unloading of bins on both sides
of the corridor.
The system described will detect objects, such as personnel,
boxes, crates, or other items that accidentally come into the path of
the vehicle.
One requirement for the system was that the detection zone cover
a large area, have sharply defined edges, and have limited range. As
a result, a simple optical beam breaking alarm, or ultrasonic intrusion
type alarm could not be used.
The detection zone required is shown in Figure 1. Since there is
no likelihood that an obstruction would be suspended from the roof of
the corridor, with no support from floors or walls, the detection zone
can attain the "U" shape shown.
In order to limit the range of detection to approximately 12 feet,
so that two vehicles might operate in the same corridor without
mutual interference, the one horizontal and two vertical portions of
the detection zone were composed of pairs of illuminators and detec-
tors, separated by the width or height of the vehicle so that only an
Figure 1. Desired Detection Zone
object lying in both the illumination beam, and within the receiver
acceptance cone, would be detected. The pattern of overlapping areas
to form either the horizontal or vertical portions of the detection zone
is shown in Figure 2.
In Figure 2, T , R,, T£, R? , etc, represent transmitter-receiver
pairs operating with different pulse repetition frequencies (PRF), so
that R_ will respond only to a target illuminated by T , and not by" 2
T or T~. An additional three transmitter-receiver pairs fill in the
zone denoted by T'R'. A block diagram of a transmitter-receiver
pair is shown in Figure 3.
If the detection zone depth could be extended to 2 1 feet, as shown
in Figure 2, the signal processing and phase-locked-loop detector
could be common to three transmitter-receiver pairs, thereby
reducing cost and size of the system. The T'R' pairs still must
operate on different frequencies, as must the transmitter-receiver
pairs composing the other two segments of the overall detection zone.
A near infrared GaAs laser diode was chosen for the illumination
source, because of its low cost, small size, and relatively high power
output (as compared to LED's). Power supply requirements for the
transmitter are modest, requiring only 1.5 watts average power (300
volts at 5 ma). In addition, a -15 volt source is required for the
transmitter, which is obtained from the receiver power supply.
ajDiqs^
10
0)4->-l-J
t-<
ato
G
H
cPhoto-
detector
0)
en•—
I
Oh
0)
Mbe
H
1
T
Pre-
amplifier
Jr
Power
Supply
Signal
Processor
and
Alarm
j-i
b0nj
• i-i
QMoo
1—1
>• 1-1
CD
V0)
I
CD
CO
uH
cu
DO
11
The receiver employs a PIN photodiode for a detector, and, with
the exception of the first stages of the receiver, integrated circuits
were employed. The entire receiver operates from a dual + 15 volt
supply, which draws 25 mA per side.
12
II. DETAILED SYSTEM DESCRIPTION
A. TRANSMITTER
1. Laser Pulsing Considerations
The laser employed in this system is a Laser Diode Lab-
oratories Model LD-22. This diode is a gallium arsenide injection
device which emits coherent infrared radiation when suitably pulsed.
The diode is built as a hetero-structure, which differs from older
diffused junction devices in that three distinct layers: n-type gallium
arsenide, p-type gallium arsenide, and p_type gallium aluminum
arsenide compose the diode structure. The hetero- junction is formed
at the interface of the p-type gallium arsenide, and the p-type gallium
aluminum arsenide, which serves to confine the injected electrons and
reduce reabsorption. These devices are more efficient than diffused
junction devices since the current at which lasing action occurs (thresh-
old current) is lowered. Radiation is along the axis of the hetero-
junction and is proportional to the forward current of the diode.
Peak Power: 6 Watts
Threshold current: 8 Amperes
Peak forward current: 25 Amperes
Table I.
13
The laser diode can only be operated in a pulsed mode, since
current pulses of more than 200 nanosecond duration would overheat
the junction and destroy the device. Specifications for the LD-22
limit the PRF to 5MHz, and pulse width to 200 nanoseconds.
A more important consideration is the shape of the current
pulse. Threshold current increases with junction temperature, as
shown in Figure 4.
If the pulsing current waveform does not have a fast rise
time, the current flowing below threshold level will not cause lasing
action, but will only heat the junction, thereby raising the threshold
level and decreasing efficiency.
2. SCR Pulsing Circuit
In order to generate a current pulse of sufficient magnitude,
and which has a fast rise-time, a capacitor discharge circuit employ-
ing a fast SCR was used. The circuit used is shown in Figure 5. The
SCR switch is an RCA 40768. In pulse circuits of this type, the SCR
is operated in an unconventional mode, i. e. , the duration of the anode
current pulse is much less than 1 microsecond, (nominal turn-on
time for most SCR's). In the conventional mode, turn-on impedance
is a fraction of an ohm. However, in the laser pulser, the capacitor
discharges in less than one microsecond, so that the SCR does not
completely turn on, and the anode-to- cathode impedance can attain a
level of several ohms during the conduction period. This impedance
considerably increases the voltage required to p\ilse the laser at .he
14
c<u
uu2>
U
•—
(
O
CO
U^i
HT3CD
N
O
1. 8
1.6
1.4
1.2
1.0 —
100
Temperature ( C^
Figure 4. Normalized Threshold Current
vrs. Junction Temperature
15
(J
u
aauH
u
rH
16
required current levels. The laser diode exhibits a forward resistance
of much less than one ohm. The only other resistive element of the
pulsing circuit is the current monitor resistor which was selected at
3 ohms.
Reactance in the pulsing circuit, from lead inductance and
wiring capacitance, can cause ringing in the current pulse. The SCR
turn off time is much longer than the ringing time, and voltage trans-
ients in excess of 3 volts of reverse bias can destroy the laser diode.
Excessive wiring capacitance will reduce the slope of the current
pulse, thereby reducing laser efficiency because of junction heating.
Selection of an SCR cannot be made on the basis of published
data alone. These data give characteristics of the SCR in a gross
sense, and do not indicate performance in high speed pulsing applica-
tions. Recommended selection is on the basis of a "use test"; the
RCA 40768 can be depended on to fulfill the requirements for laser
pulser service. Reference 8 claims the RCA GA-201 will work in a
laser pulser circuit with the high voltage supply requirement reduced
to 100 volts. The GA-201 was not available for use, so the 40768 was
used.
Circuit layout in the pulsing circuit is critical. Resistance
in the pulsing circuit, which is composed of SCR conduction resistance,
laser resistance, current monitor resistance, control diode (1N914)
resistance, will all tend to dampen out ringing. If the sum of these
17
resistances is too low, negative voltage overshoots can destroy the
laser; if the resistance is too large, pulsing efficiency is reduced.
Diode D also protects the laser from reverse bias damage.
3. Trigger Circuit
The 40768 SCR requires a trigger pulse of 10 ma, with a
duration of at least 1 microsecond. The gate to cathode impedance
is low, so the trigger source must be obtained from a low impedance
source. A simple common-emitter amplifier was constructed so that
a readily available pulse generator (GR 1217-C) might be used. In
the final design, trigger pulses are obtained from the phase locked-
loop tone decoder. A description of the required waveshaping to obtain
these pulses is described in section II. C. 4.
4. Charging Circxiit
The capacitor (C ) which delivers the current pulse to the
laser must be re-charged between successive pulses. However,
before the capacitor is recharged, the SCR must be turned off, other-
wise current would flow continuously through the laser after the initial
pulse. The holding current for the 40768 is nominally 25 ma. Resistor
R of Figure 5 limits the steady-state current from the 300 volt supply
to 8.5 ma, which is well below the holding current. Thus, the SCR
is forced to open once the currentfalls below 25 ma.
Diodes D. and D?
, due to their forward voltage drop of
approximately 1.2 volts, hold Q, in a cutoff state as long as current
flows through the SCR. When the SCR opens, Q^ is then biased into
saturation and the capacitor C charges through R . Since the R Ctr Z ** P
time constant is 86 microsecond, much less than the pulsing period of
1 millisecond, C charges to the full 300 V before the next pulseP
occurs.
Pulse current control is obtained by varying the high voltage
supply, V . For full rated output of the LD-22, V is set to 300 volts,s s
Reference 7 shows alternate methods of adjusting pulse current.
Waveforms of iD and v are shown in Appendix A.
5. Optics
The radiated emission of the LD-22 is confined to a cone of
15 at the 3 dB points. Since it was desired to shape the beam into a
fan with minimum vertical dispersion, and 10 horizontal dispersion,
a cylindrical lens of 25mm focal length was employed to collimate the
beam vertically, and the horizontal plane was masked with an aperature
to the 10° beam width. Vertical collimation was expected to give a
3 41/R response to the system rather than a 1/R response. A diagram
of the lens system is shown in Figure 6.
TRANSMITTER
Laser-j^.]}
•^^o VerticallyCollimated Ream
?Lens
Figure 6. Transmitter Beam Optics
19
6. Safety Considerations
Reference 1 states that the maximum safe light intensity of a
pulsed laser with very short pulse duration, and operating in the near
infrared (where the light is focused on the retina of the eye), is 0. 07
2joules /cm . With a laser output of 5 watts, the maximum power den.
2sity at the surface of the cylindrical lens is 11. 94 watts /cm . Convert-
ing this to an energy density, by multiplying by the pulse duration,
yields a density of 2. 388 x 10 joules/cm , which is well below the
maximum safe level. The laser should therefore present no safety
hazard to personnel.
B. RECEIVER
1. Optics
The laser pulses reflected from an obstruction in the detection
zone were focused on the detector cell by means of a condensing lens
onto a translucent mylar integrating surface. This is shown in
Figure 7.
The PIN photodiode used, an HP4203, has an acceptance
angle of 90 . Since the desired field of view of the receiver ranges
o ofrom 18 to 27% . considerable receiver aperature gain could be
realized by optical focusing as opposed to simple masking of the
receiver aperature. The lens employed is a 30 mm diameter, 55 mm
focal length (f/1. 83) lens.
20
en
\
f-\
en
Ccu
X."
v /\ /
)[
en
u.1-1
-i-)aO
CD
>•r-l
CD
U<u
0)
• r-l
21
The integrating surface was masked to provide the desired
field of view. With a 17 mm width of the integrating surface, field
oof view is 18 , as shown in Figure 7. This is the field required for
R.. For R?and R , wider integrating surfaces are required. Vertical
masking height is relatively unimportant, since the transmit beam is
collimated in the vertical plane, but should be as large as practical
in order to maximize optical gain.
The optical gain of the receiver can be expressed by the
relation:
G =r
Ai
Ad
where
G = Receiver optical gain
A. = Lens aperture area
A, = Photodiode aperture surface
2For a 30 mm diameter lens, and photodiode area of 0. 2 mm ,
706. 86G = -3534.3 ^ 35.5dB+ p (dB)r ~
P0.2
Since p< 1, the gain will be less than 35. 5 dB. The value
of P depends on many factors, including reflection and transmission
qvialities of the lens system, and losses due to masking. An empirical
value of p is given in section III. A. 3.
2. Photodiode Considerations
The PIN photo diode detector (HP 4203) is a silicon planar
device sensitive to visible and near-infrared radiation. These d
22
exhibit response speeds of less than 1 nanosecond, and have a low dark
current (5000 picoamp) which allow detection of low light levels. The
quantum detection efficiency is constant over 6 decades of light intensity,
which provides a 60 dB dynamic range. All of these features make this
type of device especially suitable for application in a system to detect
short duration laser pulses in a high ambient light level environment.
3. Impedance Converter
The PIN photodiode can be considered a current source with
a source impedance of 22 Megohms. . It is therefore essential to operate
the photodiode into a very high impedance load. A FET source follower
circuit was used, shown in Figure 8.
An equivalent circuit for the photodiode and impedance
converter is shown in Figure 9. Where i is the photo current, i is
I s ^5*. in Cs
R,
Vy\—
©
Output
Figure 9
Photodiode and Impedance ConverterEquivalent Circuit
23
>in
Hi'
H00
(M
>ini—
i
i
©
©3a,
oo
3'
\ oo #—*—'WV
oo
M-l
m
CO
^
<u-l-J
Jh
D>
O
S s
<u oa, CO
a u
a; £
.2 ao £
£ Oh
co a0)
pi
•r-l
24
the noise current of the photodiode, i is the dark current, R is ther s
equivalent source resistance of the photodiode (22 megohm), R. is the
input resistance of the impedance converter, C is the shunt capacit-
ance of the photodiode and circuit capacitance, v is the noise voltage
of the FET, and v is the voltage response of the impedance converter.
The output voltage is the sum of v and v :
v = v + v__out n P
whereR R: ns in
!i + i + i
p • p n r' R s + Rin
For relative by high signal intensity:
and
1p>> xn
1*>> l
P r
The input impedance of a MOSFET is nearly infinite, so
R // R. e R , ands /' in s'
vD a i RP p s
which, in this case, is 22 microvolts per picoamp.
4. Preamplifier
The preamplifier is a simple, two stage section employing
low-noise transistors. The first stage is a common-emitter configura.
tion, and is only remarkable in that the input impedance is higher than
usual, so that the impedance converter stage is not loaded down. The
second stage is a directly coupled emitter follower, which serves to
25
buffer the preamplifier from the signal processing circuits, described
in the next section.
5. Power Supply-
Power requirements for the receiver section were primarily
determined by the bias requirement of the PIN photodiode, which needs
20 volts to 50 volts of reverse bias for proper operation. The signal-
processing section required a dual + 15 volt supply. In order to use a
single power source, 30 volts was selected for bias voltage, and the
entire receiver operates from the dual + 15 volt supply. Decoupling
of stages was accomplished with the IKohm resistors, and 10 uf
capacitors.
C. SIGNAL PROCESSING SECTION
1. General
The function of the signal processing section is to extract
only that portion of the receiver preamplifier output signal caused by
reflection of the laser output from an obstruction in the detection zone.
Primary design considerations for the signal processor were:
1. Response to weak signals, where signal-to-noise
ratios could be expected to be as low as 3 dB.
2. A high degree of interference rejection, where
interference sources could be expected to be fluorescent and incan-
descent light.
26
3. Reject signals reflected from non-obstructive targets,
e. g. , dust and smoke.
4. Rejection of signals caused by reflections from the
target from transmitters other than the one synchronized to the
receiver.
5. Employ a threshold detector that would provide a
positive alarm-no alarm output.
In order to meet these objectives, it was decided to use
bandpass active filters, that would satisfy requirements (1), (2), and
(4), and to use a phase-locked-loop tone decoder to satisfy require-
ments (2), (3), and (5).
A block diagram of the signal processor section is shown in
Figure 10.
2. Active Filter
Since the frequency spacing of the PRF of each transmitter-
receiver pair could be separated by several hundred hertz, the primary
restraint on the filters was to provide narrow enough bandwidth to
reject interference caused by incandescent and fluorescent lights.
Incandescent lights provide little or no interference, as their luminous
output is very nearly constant, with only a slight amount of 120 Hz
modulation. On the other hand, fluorescent lights create a consider-
able amount of interference signals in the frequency range of the
transmitter-receiver pairs. The detector response signal to the
fluorescent lights has a fundamental 120 Hz component, but, because
27
<D DC
0)V-i
o
O
a)
m•r-l
Q
w t-i
V 0)
m *->+J•r-l
aw
Jh a<D
RjDC u
• iH H ifU nt
-J
H
Jh
DC
•i-i
P
oot—
i
mJn
Owen
oo
cDCrHC/3
DC
28
of the non-linear characteristics of the gas -discharge tube, and the
noise produced by the gas -discharge, the detected luminous output has
frequency components extending well past 2000 Hz, and spaced every
60 Hz. Figure 11 shows a portion of the spectral content of 96"
fluorescent. lights.
It was therefore necessary to select operating frequencies
centered between the spectral lines of the fluorescent lights. For
example, an operating frequency of 1050 Hz lies between the 1020 Hz
and 1080 Hz interference frequencies. Rejection of these interference
frequencies was only on the order of 2 dB, but the tone decoder des-
cribed in the next section provided sufficient rejection of the inter-
ference frequencies. The filters did extract the fundamental frequency
of the input pulses, and provided a gain of 43. 5 dB at 1050 Hz. (See
Appendix B). The filters are shown schematically in Figure 12.
,01^/tf
-\{r14.7 Mrt
/\Mr—100 KfL
©—AA/VInput
10 Kn
Q +15 V
Output
Figure 12
Active Filter Schematic(Single Stage)
29
r-
-10
_20
a
C
&N
O
PJ
C0)-J-J
coU
-30
.40 -
50
o
& -60
70
^--
Oroo
900
Figure 1 1,
NXo00o
1000 13001100 1200
Frequency (Hz)
Portion of Spectral Content of Detected Waveform of 96'
Fluorescent Lights
30
3. Phase-Shifter, Buffer, and Limiter
The tone decoder requires that the input signal be 90 out of
phase with the reference oscillator contained within the tone decoder '
in order to achieve lock. In an open loop system, the reference
oscillator can shift frequency (and phase) to lock onto an input signal
that is within the capture range. However, in this system, since the
input signal is derived from the reference oscillator, and since all
circuitry preceding the tone decoder provides either 0° or 180° phase
shift, the reference oscillator would shift in frequency to a point where
the active filters provided a 90° phase shift, and then achieve lock.
Since the filter phase characteristics were such that this frequency
was on one side or the other of the amplitude response curve, a loss
in overall system sensitivity resulted.
In order to eliminate this problem, an operational amplifier
phase shifting circuit was employed, which provided the required 90°
phase shift. The circuit is shown schematically in Figure 13.
Input
10 KAO—5—V^/V-
10 K<^Ph'a
> i
PhaseAdjust.
10 KO
GainAdjust ol^f
7U1 ">-H^THf-10 Kn-
0-15 V 1 KO.
Output
2ZX(2) 1N91U
o
Figure 13
Phase-Shifter, Buffer, and Limiter Schematic
31
In addition to providing phase shift, this stage served to buffer
the output of the active filter from the tone decoder, which has a low
input impedance.
To prevent excessive input signal amplitude to the tone decoder,
the two diodes were employed to provide hard limiting of the input
signal. Signals in excess of 1.2 volts peak amplitude are clipped.
The gain control, R, provides a means of controlling the
sensitivity of the receiver.
4. Tone Decoder
The final stage in the detection process employs a Signetics
567 tone decoder. This device was chosen because of the narrow
capture bandwidth available, response to low amplitude input signals,
and simplicity of external circuit design. The 567 also provided a
convenient source for the transmitter trigger pulses. The tone
decoder stage is shown schematically in Figure 14.
I v U."
+lt Alarm Output
TriggerOutput ">}!
Rx - 20 K/l*-h r
2- io Kn.
R-> - 390nR? - 820a
C-, - 0.1 /cf
C?
- 2S/*fc3
- $0 /*f
C, - 0,1 yUf
Cg - 0.001 /£
D-j^ - 1N755A (7.£V Zener)
Figure 14 -zbr
Tone Decoder Schematic
32
The circuit elements R, and C, determine the free-running
frequency of the reference oscillator: R. is a 10 turn trimpot and
was adjusted for a free-running frequency of 1050 Hz. Capacitors
C and C control lock-on time delay, and capture bandwidth. They
also interact to control the centering of the capture bandwidth (skew)
about the free- running frequency. With the values specified, the
capture bandwidth is skewed toward the lower frequencies; this was
done in order to provide greater rejection to the 1080 Hz component
of fluorescent light interference, which was 36 dB stronger than the
1020 Hz component. (See Figure 11).
Lock-on time was set at approximately 0. 1 second (i. e. , the
tone decoder must receive about 100 cycles at the input before lock-on
occurs). This was done in order to prevent false alarm conditions
resulting from non-obstructive reflectors, such as dust and smoke.
Reflectors such as these provide a scintillating reflection of very
short duration.
The tone decoder can sink up to 100 ma when a lock condition
occurs, and this sink can drive whatever suitable alarm circuit is
desired. Since this portion of the tone decoder is compatible with
most logic circuitry, the several tone decoders could drive a central
multi-inpxit OR gate that set an alarm condition (i. e. , stop the
vehicle) wherever any receiver sensed an obstruction.
The reference oscillator is the source of trigger pulses for
the transmitter. This scheme has the advantage of producing
33
synchronous detection, since the signal that the tone decoder will
respond to is generated by the tone decoder itself. Thus, any frequency
drift in the reference oscillator will not degrade system performance,
as long as the drift does not exceed the passband limits of the active
filters. In addition, the input signal must be 90 (i5 ) out of phase
with the reference oscillator. This fact reduces the susceptability to
undesired signals, in that both the frequency and phase of the input
signal must be Within very close limits to produce an alarm condition.
The reference oscillator output is a five volt peak-to-peak
square wave, and it was necessary to differentiate this signal to pro-
vide negative spikes of sufficient magnitude to drive the transmitter.
This was accomplished with a simple RC differentiator circuit.
34
III. TEST RESULTS
A. RANGE TESTS
With a transmitter receiver pair arranged in the configuration of
Figure 2, operating at a PRF of 1050 Hz, a positive alarm condition
2was achieved with a white cardboard target of 1 ft area at a maximum
range of 16 feet. The edges of the detection zone were sharply defined,
with an edge definition of less than one inch at the 16 foot range.
In order to more fully investigate the performance of the trans-
mitter and receiver optics, and to develop a range equation for the
overall system, more extensive tests were conducted.
1. Reflection Tests
For these tests, a target of known area was positioned at
various ranges from the transmitter. Range from the transmitter
2and receiver was the same. The target was a 1 ft white cardboard
square. Return signal strength was measured at the output of the pre-
amplifier. Since the photocurrent of the photodiode is directly pro-
portional to the radiant power density incident on the photodiode,
v -
relative signal strength was computed as: 10 log . Figure 15v ref
shows the response curve of the target described. The reference,
v £, was arbitrarily chosen as the output voltage with the target at a
range of two feet.
35
Slope of 1/RResponse
Slope of
l/R-* Response
J JL 1 I I
3 4 56789 10 15
Range (feet)
Figure 15. Range Response of a 1 ft Target
20
36
The portion of the response curve from two feet to nine feet
does not follow the 1/R response that was expected. This was caused
by aperture limiting the portion of the target illuminated by the trans-
mitter and seen by the receiver: as the range increased, effective
target area also increased.
At ranges greater than nine feet, aperture limiting ceases,
3and the response follows the expected 1/R~ form.
2. Target Response
Tests of response to various target materials were made.
Results are tabulated in Table II. The "half power angle" is the angle
of the target (from the normal to the transmitter) at which the reflected
power is reduced by one-half. For an isotropically reflecting surface,
Lambert's Law would indicate that this angle should be 60°. Since
this angle is less in all cases, none of the targets tested are purely
isotropic reflectors.
3. Direct Line-of-Sight Tests
These tests were conducted in four phases: no optics,
transmitter optics alone, receiver optics alone, and with all optics
employed. They were expected to show the performance of each
section of the optics employed, as well as to demonstrate the potential
of the transmitter and receiver for uses other than obstruction
detection.
The transmitter and receiver were aligned in a straight
path line-of-sight, and again, preamplifier output was taken as a
37
Target Reflection Half PowerMaterial Loss (dB) Angle (°)
Mirror
Plexiglas -2.36 1
AluminumSheet -6.8 3.5
Plywood -29. 3 42.5
Brown Paper -29. 8 45
WhiteCardboard -30.3 40
Light ColoredCotton -30. 3 45
Medium ColoredWool -30.3 50
GreenCardboard -30.9 40
Dark ColoredCotton -31.5 45
Table II.
Relative Reflection Characteristicsof Various Targets
38
measure of response. The results are shown in Figure 16, and
demonstrate the gain realized from the optics.
At a range of 100 feet, the transmitter optics provide a gain
of 15 dB, the receiver optics provide a gain of 8 dB, and all optics
combine to achieve a gain 22. 5 dB.
Without optics, the range response is slightly better than
4the 1/R form expected. This can be attributed to the fact that the
laser output is somewhat collimated as it exits the junction, and can-
not be considered an isotropic source.
Figure 16 also shows the effect of saturation of the pre-
amplifier stage with signals greater than about six volts (-5 dB), and
the fact that the photodiode. saturates at light levels greater than that
required for a 20 volt (0 dB) output.
In section I. B. 1. , the maximum theoretical gain of the
receiver optics was calculated as 35.5 dB. Test results show a gain
of 8 dB for the receiver optics. The 27. 5 dB loss is attributed to the
integrating surface. This loss is composed of two factors: attenuation
and reflection losses of the mylar surface, and re-radiation effects
from the surface. Attenuation and reflection losses could be measured
directly, and were shown to amount to -4.2 dB. The re-radiation
effect makes up the remaining loss. If each element of the integrating
surface can be considered a point source, the solid angle from that
point to the photodiode surface subtends a solid angle of only
31. 39 X 10" steradians. Thus the theoretical loss from each point is
39
OTg[
(ap) ^uod S oH 9a„ BT9H
40
1.39x_i0l3
=2 . 2xl0 -4 ^ -43.4 dB.2
The actual re-radiation loss is 23. 3 dB, and the difference can be
explained as arising from the fact that the target image on the surface
is of finite size, which the photodiode integrates, and that some reflec-
tions from inside the aluminum mount for the integrating surface land
on the photodiode surface.
B. SPURIOUS RESPONSES
1
.
Ambient Light
With the system operating in a high level fluorescent light
environment, no false alarm conditions were encountered, as long as
the operating frequencies were set between the harmonic frequencies
of the fluorescent light signal.
By using a high-intensity strobe lamp, false alarm and jam-
ming (non- response to target) could be obtained, but this type of source
could not be expected to be encountered in normal service.
2. Dust and Smoke
No false alarm conditions were encountered when dense
tobacco smoke was blown into the detection zone. A type of "chaff",
small paper discs from a three-hole punch, was sprinkled into the
detection zone, with no false alarm condition resulting.
41
3. Electromagnetic Compatability
The pulsing circuit emits electro-magnetic radiation fields
that could be detected with a portable AM radio at distances of up to
10 feet. A plot of this radiation, measured with an AN/PRM-1 field
intensity meter is shown in Figure 17.
Attempts were made to make the receiver respond to radio
frequency signals. A signal generator with a 0. 5 watt output was
coupled to the receiver with a three inch loop of wire located six inches
from the photodiode. No response was noted with the signal generator
operating from 10 MHz to 250 MHz. These tests were conducted with
both CW output, and with the output modulated at the operating frequency
of the tone decoder.
4. Adjacent Transmitter Interference
A target was illuminated by two other transmitters, operating
at 1110 Hz and 1170 Hz. These represent the closest frequency spacing
that could be expected to be employed (again, centering the operating
frequencies between fluorescent light harmonics). No response to
either transmitter by the tone decoder was noted.
42
Figure 17
5 10
Frequency (MHz)
Radiated Interference Frequency Spectrum
43
IV. CONCLUSIONS
A. RANGE RESPONSE
3It was assumed during the initial design phase that an 1/R
response could be expected since the transmitter beam was collimated
in one dimension. Test results showed that the range response followed
. 3a 1/R slope very closely.
The maximum range required was 12 feet, although minimum
target size was not specified. To determine the minimum target size
detectable at any arbitrary range, a range equation of the following
form can be used:
R 3= K°
Where
K = a constant based on transmitter power, gain of transmitter
and receiver optics, and minimum signal_to-noise ratio for valid
alarm.
= target effective area, which depends on reflection coefficient
and angle of incidence and reflection.
2
Using a 1 ft white cardboard square as a target of ° =1, the
2minimum effective target area is 0. 422 ft for response at a range of
12 feet.
44
B. TARGETS
The effective target area is a very difficult factor to predict:
much like the effective radar cross -section, it is statistical in nature.
Targets with a high reflection coefficient (e. g. , mirrors, smooth
metal, plexiglas) make poor targets since the angle of incidence and
reflection is very critical. Targets of much lower reflection co-
efficient make better targets, since the angle can vary considerably
with much less loss in effective area.
C. ENVIRONMENTAL CONDITIONS
It was intended to design a system that cotild operate in high
ambient light conditions. This goal was achieved, largely due to the
very narrow capture bandwidth of the phase-locked-loop tone decoder,
which is the heart of the synchronous detection system employed.
The use of integrated circuits for the active filters and tone
decoder allow the use of these types of circuits in a system of small
size with modest power requirements.
45
APPENDIX A.
Photographs of Various Voltage and Current Waveforms
Preamplifier andFilter Output,
High S/N Ratio.
Horiz: .2 ms/Div.
Vert:
Upper: . 2V/Div.Lower: lV/Div.
Preamplifier and '
Filter Output,
Medium S/N Ratio.
Horiz: 2 ms/Div.
Vert:
Upper: . 05 V/DivLower: . 2 V/Div
46
Preamplifier andFilter Oiitput,
Low S/N Ratio.
Horiz: .2 ms/Div.
Vert:
Upper: . 02 V/Div.Lower: . 1 V/Div.
•SCR Gate Voltage
Horiz: .2 ms/Div.
Vert: 2 V/Div.
47
Capacitor Voltage,vcP
Horiz: . 2 ms/Div.
Vert: 50 V/Div.
Pulsing Current, i .
P
Horiz: . 5 us/Div.
Vert: 10 V/Div.
48
Appendix B
Active Filter Frequency Response
$f Of S£ 0£ 9Z OZ SI
(gp) uttsq J3T[ij:
01
49
BIBLIOGRAPHY
1. Department of the Navy Publication NAVMED P_ 5052-35,
Control of Hazards to Health from Laser Radiation,
p. 3-4, 24 February 1969.
2. Eleccion, M. , "The Family of Lasers: A Survey" Spectrum,
p. 37-39, March 1972.
3. Grebene, A. B. , "The Monolithic Phase-Locked Loop - AVersatile Building Block", Spectrum, p. 38-49, March 1971.
4. Hewlett Packard Application Note 915, Threshold Detection of
Visible and Infrared Radiation With PIN Photodiodes.
5. Hewlett Packard Application Note 917, HP PIN Photodiode .
6. Hudson, R. D. , Infrared System Engineering, Wiley, 1969-
7. Radio Corporation of America Application Note AN-4469,Solid-State Pulse Power S\ipplies for RCA'GaAs Injection
Lasers, DeVilbiss, W. F. and Klunk, S. L. , February, 1971
8. Radio Corporation of America Application Note AN-4741,A High Speed Pulser for Injection-Laser Diodes, De Vilbiss,
W. F. and Klunk, S. L., p. 5, September, 1971.
50
INITIAL DISTRIBUTION LIST
No. Copies
1. Defense Documentation Center 2
Cameron Station
Alexandria, Virginia 22314
2. Library, Code 0212 2
Naval Postgraduate School
Monterey, California 93940
3. Assoc Professor G. L. Sackman, Code 52Sa 12
Department of Electrical EngineeringNaval Postgraduate SchoolMonterey, California 93940
4. LT George Alfred Burman, USN 1
Naval Electronics Systems Command (PMS116)Washington, D. C. 20360
5. Mr. Lane Lee 1
Hewlett Packard Corporation620 Page Mill -Road
Palo Alto, California 94304
6. Mr. Jack Mattis 1
Signetics Corporation811 E. Arques AvenueSunnyvale, California 94086
51
UnclassifiedSecurity Classification
DOCUMENT CONTROL DATA -R&D[Security cles nfic mtion ol title, hodv ol abstract and indexing annotation must be entered when the overall report Is classified)
1 originating AC Ti VI T Y (Corporate author)
Naval Postgraduate School
Monterey, California 93940
2«. REPORT SECURITY CLASSIFICATION
Unclassified2b. GROUP
3 REPORT TITLE
An Optical Object Detection System for Sensing Obstructions to LowSpeed Vehicles
4 DESCRIPTIVE NOTES (Type of report and.inclusive dates)
Master's Thesis; June 19729. au THORiSi (First name, middle initial, last name)
George A. Burman
6 RLPOR T D A TE
June 1972
7«. TOTAL NO. OF PAGES
53
7b. NO. OF RE FS
8a. CONTRACT OR GRANT NO.
b. PROJEC T NO.
9a. ORIGINATOR'S REPORT NUMBER(S)
9b. OTHER REPORT NOISI (Any other numbers that may be assignedthis report)
10. DISTRIBUTION STATEMENT
Approved for public release; distribution unlimited.
11. SUPPLEMENTARY NOTES 12. SPONSORING MILITARY ACTIVITY
Naval Postgraduate SchoolMonterey, California 93940
13. ABSTRACT
An object detection system to sense obstructions in the path of low speedvehicles is described. The system uses a pulsed GaAs diode laser as anilluminator, and a PIN photodiode as a detector. Reliable detection of objects
with an effective area as small as 0. 5 ft was achieved. A signal processorusing active filters and a phase-locked loop tone decoder was employed for
both phase and frequency rejection of undesired signals, and detection of
objects under high ambient light conditions.
FORMI NOV 6S
S/N 0101 -807-681 1
1673 (PAGE n
52UnclassifiedSecurity Classification
i-3l«oe
Unclass ifiedSecurity Classification
KEY WO RD!
Photodiode Detector
Optical Obstruction Detection
Active Filter
Phase- Locked Loop
GaAs Diode Laser
Laser Safety Requirements
PIN Photodiode
DD,?»?„1473 <back UnclassifiedS/N 0101-807-6921 53 Security Classification
Thesis 135765B88368 Burman
c.l An optical object de-
tection system for sens-
ing obstructions to low
speed vehicles..
ThesisB8836J
fer.j
135765Burman
An optical object de-
tection system for sens-
ing obstructions to low
speed vehicles.
thesB88368
An optical object detection system for s
3 2768 001 02091DUDLEY KNOX LIBRARY